Miserable day...losing faith

[cjcj1949]

There is nothing special about the position at switch on, the moment you make a rudder input, it is forgotten.

The normal small rudder inputs you would make on take-off cancel any 'error' immediately.

I think my original statement is correct. The gyro will have a position that it thinks it should be in and as soon as it is able to turn it will try to move to that position. If you give an input on rudder that will be a yaw rate input, as it is always, and this will modify the desired position. It won't reset it. With good gyros and good pilots the difference is negligable. With older gyros a 180 degree turn on take off stays in the memory.

I'd like to see a video of this problem if possible. Also you can learn a lot by just turning the heli on the ground in HH. The rudder servo should come back to the same position as the heli is turned back to the same heading. I agree with Mark that this sounds an impossible fault. That is not that unusual though and will be either a misunderstanding on our part of the explanation, or something lacking in our understanding of what is going on, or something really silly. It sound very interesting though.

Thinking about Mark's comment, it is possible that there are two ways of handling new input. The gyros I've seen do not go back to the mid point as soon as a small input is given. As far as I can see it would be quite reasonable to do so. It might be obvious or it might not. Resetting the integrator every control movement sounds reasonable, but I'm not sure what that would do to handling in wind as you would be interfering with the control loop's correction.

[Mark_T]

If you give an input on rudder that will be a yaw rate input, as it is always, and this will modify the desired position. It won't reset it.

That doesn't match my experience and thinking from a software point of view it wouldn't make a lot of sense to try to track against a fixed start position as that would ensure you were always fighting against accumulated errors.

You'd always want the yaw input to be relative to your current heading and reset to the point where the yaw input stops, surely?

Just by way of example, the way I set-up on the bench is normally 180 degrees to the position out at the peg. I always put in a little left and right rudder to check all is OK during spool-up, never yet seen any tendency to try and do a 180 during take-off.

[cjcj1949]

If you make a small yaw correction the rudder servo doesn't go back to centre. I'm not sure why as I said. If it did reset the integrator there wouldn't need to be separate sequences of operation to reset it. The easiest way is to be in rate mode until take off. Giving a small amount of left and right doesn't centralise it but it does get in away from the edges. If you have it in the right position to begin with then a small flick either side won't make much difference. I have noticed that each gyro does its own thing. I'm not sure about the software argument as it's the control algorithm that does the job and different error control methods would be needed depending on the sensor errors within the operating envelope. Whatever the theory says it doesn't always survive testing.

You are right that a some control input will minimise the error, but you can't really claim that it gets rid of the error unless you know that there is an error there. My 180 degree experience was on a "competition standard" Twister Storm Gyro from JP that had appalling temperature performance. That was how I discovered how they worked and how to deal with a bad gyro. My latest gyro is under £10 and a mems unit. (Assan 250)